Particle-reinforced aluminum matrix composites are widely used in aerospace and rail transportation fields due to their excellent specific strength and specific stiffness. However, their further development is limited by brittle fracture and the strength–toughness trade-off caused by the introduction of reinforcements. Traditional quasi-static characterization methods are inadequate for capturing the dynamic damage evolution of materials under load. To address this challenge and reveal the dynamic fracture mechanisms of Particle-reinforced composites, this study developed an in situ dynamic continuous tensile X-ray computed tomography (CT) method with a temporal resolution of up to 5 Hz at the BL16U2 beamline of the Shanghai Synchrotron Radiation Facility. This approach enabled 4D (3D space and time) observation of the entire fracture process of a 12 vol.% Ti2AlC/Al composite. By combining deep learning and ambient occlusion (AO) algorithms, the challenges associated with accurately segmenting reinforcements and microcracks in low signal-to-noise ratio dynamic images were successfully addressed. The findings revealed that: (1) Damage initiation exhibited significant size dependence, with microcracks preferentially nucleating inside larger and agglomerated Ti2AlC particles; (2) Damage evolution followed a three-stage characteristic of “multiple-site crack initiation–matrix connection–localized coalescence”; (3) At the critical point of failure, isolated microcracks rapidly coalesced by the Al matrix cracks, leading to a sharp decrease in the number of cracks and a drastic increase in the volume proportion of the main crack and ultimately triggering unstable fracture. This study elucidated the dynamic fracture mechanism of 12 vol.% Ti2AlC/Al composite and demonstrated the unique advantages of sub-second high spatiotemporal resolution CT in capturing nonlinear failure behaviors in materials.